Kinetic study of ethanol dehydrogenation to ethyl acetate promoted by a copper/copper-chromite based catalyst

Article abstract of DOI:10.1016/j.cattod.2012.02.054

A kinetic study of the ethanol dehydrogenation to ethyl acetate on a copper/copper-chromite catalyst has been performed. The used catalyst, in cylindrical pellets, contained also alumina as a support and barium chromate as a promoter. Support and promoter have the effect of increasing the activity, the selectivity and the stability of the catalyst, as shown in a previous work. The kinetic runs were carried out in a packed bed tubular reactor, alternatively filled with 2 or 50 g of catalyst, approximately isothermal, by feeding pure ethanol together with a mixture of nitrogen and hydrogen as carrier gas. Kinetic runs have been made by changing the temperature, in the range of 200-260 °C, the pressure between 10 and 30 bar and the space time from 1 to 100 g h mol^-1. We have verified, at first, that inter-phase and intra-phase mass transfer limitations were negligible in the adopted conditions. Then, a Langmuir-Hinshelwood-Hougen-Watson kinetic model has been used for interpreting all the experimental data collected. This model corresponds to a mechanism in which the first step is the dissociative adsorption of ethanol on the surface, giving an adsorbed ethoxy group. Then, two other consecutive steps give place to respectively acetaldehyde as intermediate and ethyl acetate. This kinetic model allows a satisfactory fitting of all the performed experimental runs with a standard error below 15% for the runs performed with 2 g of catalyst and less than 12% for the runs made with 50 g of catalyst.

Full text of DOI:10.1016/j.cattod.2012.02.054

Catalysis Today 203 (2013) 202–210
Contents lists available at SciVerse ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Kinetic study of ethanol dehydrogenation to ethyl acetate promoted by a
copper/copper-chromite based catalyst
∗
G. Carotenuto, R. Tesser, M. Di Serio, E. Santacesaria
University of Naples “Federico II”, Department of Chemistry, Complesso Universitario Monte S. Angelo, Via Cintia 4, IT 80126 Naples, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
A kinetic study of the ethanol dehydrogenation to ethyl acetate on a copper/copper-chromite catalyst
has been performed. The used catalyst, in cylindrical pellets, contained also alumina as a support and
barium chromate as a promoter. Support and promoter have the effect of increasing the activity, the
selectivity and the stability of the catalyst, as shown in a previous work. The kinetic runs were car-
ried out in a packed bed tubular reactor, alternatively filled with 2 or 50 g of catalyst, approximately
isothermal, by feeding pure ethanol together with a mixture of nitrogen and hydrogen as carrier gas.
Received 8 November 2011
Received in revised form 2 February 2012
Accepted 24 February 2012
Available online 5 April 2012
Keywords:
Bio-ethanol
Ethyl acetate
Dehydrogenation kinetics
Copper chromite catalyst
◦
Kinetic runs have been made by changing the temperature, in the range of 200–260 C, the pressure
−1
between 10 and 30 bar and the space time from 1 to 100 g h mol . We have verified, at first, that
inter-phase and intra-phase mass transfer limitations were negligible in the adopted conditions. Then,
a Langmuir–Hinshelwood–Hougen–Watson kinetic model has been used for interpreting all the experi-
mental data collected. This model corresponds to a mechanism in which the first step is the dissociative
adsorption of ethanol on the surface, giving an adsorbed ethoxy group. Then, two other consecutive steps
give place to respectively acetaldehyde as intermediate and ethyl acetate. This kinetic model allows a
satisfactory fitting of all the performed experimental runs with a standard error below 15% for the runs
performed with 2 g of catalyst and less than 12% for the runs made with 50 g of catalyst.
©
2012 Elsevier B.V. All rights reserved.
1
. Introduction
different authors [1–10] and recently by us [11]. Until now, two
industrial processes were proposed [12,13]. Ethanol gives place in
one-step to ethyl acetate by dehydrogenation, occurring according
to the following overall stoichiometry:
The bio-ethanol production has grown very much in the world,
in the last decade, in particular in some countries as Brazil, for
example, because, it can be used as a renewable clean fuel with
a low environmental impact, for a partial or total substitution of
gasoline. Bio-ethanol is produced by fermentation of sugars and
the main sources are today the bagasse from sugar cane and starch
from maize. A challenge of the next future will be to produce bio-
ethanol starting from second-generation raw materials such as
ligno-cellulosic based materials that are not in competition with
food and are much more abundant and cheap than the actual
sources of fermentable sugars. Therefore, it is possible to foresee, a
large availability of ethanol, as renewable energy vector, at much
lower cost. In these conditions, bio-ethanol will become also an
interesting building block for producing chemical commodities like
ethylene, ethane, acetaldehyde, ethyl acetate, acetic acid, diethyl
ether etc. and related derivatives. In this paper, we focused our
attention on the possibility to obtain ethyl acetate directly in one-
step from ethanol. The reaction was previously studied by some
2CH3CH2OH → CH3COOCH2CH3 + 2H2
(1)
As it can be seen, from this reaction both ethyl acetate and hydrogen
are obtained as products. This route seems very promising, because,
ethyl acetate is a largely employed solvent but it could become the
raw material for obtaining acetic acid by hydrolysis and related
derivatives.
Copper is a good catalyst for promoting this reaction occurring
in relatively mild conditions. In particular, it was observed that by
operating at atmospheric pressure acetaldehyde is the main prod-
uct, but moderately increasing the pressure up to 20–30 bar the
selectivity is shifted toward ethyl acetate as main product [11].
Copper supported on alumina, for example, allows to obtain ethyl
acetate, with interesting selectivity, in the range of 80–85% [11].
Unfortunately, these types of catalysts are subjected to fast deacti-
vation probably due to both sintering and fouling [11]. Much higher
is the selectivity obtained by using copper-chromite based catalysts
that, for unsupported catalysts, can reach values of 94–95% and,
very important, the catalysts are stable along the time. At last, very
recently [11] we observed that a copper/copper chromite catalyst,
∗ Corresponding author. Tel.: +39 081 674027; fax: +39 081 674026.
E-mail address: elio.santacesaria@unina.it (E. Santacesaria).
0
920-5861/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.cattod.2012.02.054

G. Carotenuto et al. / Catalysis Today 203 (2013) 202–210
203
ethanol dehydrogenation to ethyl acetate may be described by the
following reaction scheme:
Nomenclature
r
observed reaction rate per unit particle volume
CH CH OH → CH CHO + H
(2)
(3)
3
2
3
2
3
[mol/(min cm )]
rj
C0
reaction rate [mol/(min gcat)]
CH CH OH + CH CHO → CH COOCH CH + H
3
2
3
3
2
3
2
reactant concentration at the external surface of the
3
particle [mol/cm ]
That is, the formation of ethyl acetate is the consequence of two
consecutive reactions, with the production of acetaldehyde as
intermediate step. As explained in the already mentioned previ-
ous work a high selectivity to ethyl acetate is favored by a fast rate
of reaction (3) draining the acetaldehyde concentration. This point
is very important because acetaldehyde can give place, through the
aldolic condensation, to several possible by-products lowering the
selectivity and favoring catalyst deactivation by fouling. A reason-
able complete reaction network has been proposed by Inui et al. [8]
and is reported in the simplified Scheme 1.
Our kinetic investigation starts from this scheme and from the
reaction mechanisms proposed in the literature [4] with the aim to
find the most reliable kinetic laws. According to a previous ethanol
dehydrogenation to acetaldehyde kinetic study, made by Tu et al.
[3], on unsupported Cu and Cr promoted Cu, this reaction should
be of first order with respect to ethanol and should have an appar-
ent activation energy of 12.2 kcal/mol. However, to our knowledge,
only a paper has been published concerning the kinetics of ethyl
acetate formation on Cu based catalysts, in particular on Cu/copper
chromite catalyst [4]. Colley et al. [4] have demonstrated that, in a
first step, ethanol adsorbs on the Cu component of Cu/Cr2O3 cata-
lysts, as an ethoxy specie, with an activation energy of 7.41 kcal/mol
and, successively, the adsorbed ethoxy specie, is dehydrogenated
to an acetyl specie with an activation energy of 22.50 kcal/mol.
Finally, the ethoxy and acetyl species react each other to form
adsorbed ethyl ethanoate that finally desorbs with an activation
energy of 43.06 kcal/mol. In the present paper, we have studied the
kinetics of this reaction performed on a commercial copper/copper
chromite catalyst supported on alumina and promoted with barium
chromate. In the kinetic approach, we have tested many different
kinetic models but the best fittings have been obtained by applying
a Langmuir–Hinshelwood–Hougen–Watson (LHHW) model based
on a dual site adsorption.
Rp
Deff
Fi
radius of the particle [cm]
effective diffusivity coefficient [cm /min]
2
molar flow rate of component i [mol/min]
inlet ethanol flow rate [mol/min]
0
F
E
XE,XEtOH ethanol conversion [–]
Si
Di
selectivity to component i [–]
molecular diffusivity of ethanol in nitrogen
2
[
cm /min]
εB
ꢀ
ꢁ
ꢂ
porosity of the bed [–]
bulk density of the catalyst bed [g/cm ]
tortuosity factor [–]
3
B
constriction factor [–]
mass transfer coefficient [cm/min]
reaction order [–]
kc
n
dp
v0
Ac
U
ꢃ
nCi
Aci
z
pellets diameter [cm]
volumetric flow rate [cm /min]
3
2
cross sectional area of the reactor tube [cm ]
superficial velocity = v /Ac [cm/min]
0
2
kinematic viscosity [cm /min]
number of carbon atoms in component i [–]
chromatographic area of component i [–]
dimensionless bed length [–]
W,wcat catalyst weight [g]
˛ij
stoichiometric coefficient of component i in reaction
j [–]
number of reactions [–]
axial coordinate [cm]
length of the catalytic bed [cm]
conversion or selectivities [–]
number of experimental measurements [–]
Sherwood number [–]
Nr
L
LBED
ai
N
ꢀ
Sh
ꢂi
k1
k2, k3
bi
catalytic site occupied by component i [–]
kinetic constant of reaction j [mol/(g h atm)]
2
[mol/(g h atm )]
2. Experimental
adsorption parameters [atm−1]
Kej
equilibrium constant of reaction j [–]
partial pressure of component i [atm]
adsorption constant of component i [atm−1]
2.1. Employed catalysts and pre-treatments
Pi
bi
The commercial catalyst (BASF Cu-1234-1/16-3F), employed in
the kinetic study of ethanol dehydrogenation and supplied by BASF,
is a pre-reduced copper/copper chromite catalyst supported on
alumina and containing BaCrO4 as promoter. The catalyst compo-
sition, provided by the supplier, is CuCrO /CuO/Cu/BaCrO /Al O
Subscripts
EtOH
AcH
ethanol
acetaldehyde
ethylacetate
hydrogen
4
4
2
3
EA
H2
(45:1:13:11:30% b.w.). As this is a pre-reduced catalyst, BaCrO4,
very probably, is present in a reduced form that is, as barium
chromite or BaO/Cr O . Moreover, the presence of a low amount of
2
3
CuO is due to the surface oxidation of Cu metal made by the supplier
for stabilizing the catalyst before the use. The catalyst shape is char-
acterized by cylindrical extrudates with 1.8 mm of diameter and
irregular heights of 3–5 mm. Before the use in the kinetic runs the
catalyst was submitted to a pretreatment consisting in a reduction
with hydrogen, mixed with nitrogen (H /N = 6/94 mol/mol) and
containing alumina as support and barium chromate as promoter,
increases very much the selectivity level at 98.5–99%, maintain-
ing a satisfactory activities (conversion about 65%) and showing
high stability to sintering and fouling. We have explained, in a
previous work, the reasons of this behavior also thanks to a deep
investigation on the surface characteristics of the mentioned cata-
lysts [11]. In the present paper, we have focused our attention on
the study of the kinetics of the reaction by using the mentioned
copper/copper chromite mentioned catalyst in the perspective of
the process development. As reported in a previous work [11], the
2
2
3
fed with a flow rate of 25 cm /min, keeping constant the tempera-
◦
ture at 200 C, for 16–18 h. This treatment with hydrogen reduces
CuO to Cu and about 50% of Cu2+ of the copper chromite to Cu0
according to the following reaction suggested by different authors
[4,11,14].
Cu2 + H ↔ 2H + Cu
+
+
0
(4)
2

2
04
G. Carotenuto et al. / Catalysis Today 203 (2013) 202–210
Scheme 1.
Despite the Cu2+ reduction, the spinel structure of copper chromite
is not destroyed, thanks to the formation of H inside crystalline
lattice [4,14]. In Table 1 the main characteristics of the used catalyst
are summarized.
temperature, the pressure and the ethanol space time on the
catalysts performances. Results are reported in terms of ethanol
conversions and products selectivity. The ethanol conversion is
defined as:
+
mols EtOHreacted
mols EtOHfed
XEtOH
=
2
.2. Catalytic tests
ꢀ
ꢂ
1
The catalytic tests were performed in a stainless steel packed
= 1 −
ꢁ
(5)
(
nC /nC
)(Ac /AcEtOH)
ETOH
i
i
bed reactor. A weighed amount of catalyst, in pellets (2 g) was
charged in the reactor and, before the catalytic tests, was submitted
to the already mentioned pretreatment with a flow stream of H –N
mixture (H /N = 6/94 mol/mol) of 25 cm /min for about 16–18 h.
The reaction was carried out at different temperatures in the range
While, the selectivities, determined on the basis of a carbon balance
for each component, are determined as:
2
2
3
2
2
mols producti formed
nCi
Si =
= nC
×
◦
nEtOH reacted
nCEtOH
2
00–260 C and different pressures in the range 10–30 bar. The
ꢃ
ꢄ
kinetic laws, derived from the interpretation of the runs performed
with 2 g of catalyst, have been verified by using more catalyst (50 g)
in the same reactor. The liquid ethanol reactant (Fluka 99.8% b.w.)
was vaporized at a temperature of 200 C into a pre-heater filled
with inert material (glass balls) and mixed with the carrier flow
constituted by the same mixture H /N used for the pre-treatment.
The reaction was always conducted in the presence of a hydro-
gen stream. The ethanol space time W/F, for the runs made on 2 g
nC
Aci
1 − X
i
EtOH
(6)
X
EtOH
EtOH
AcEtOH
◦
where nC and nC
bon atoms in the component i and in the ethanol fed, while Ac and
AcEtOH are the normalized chromatographic peaks areas. The mea-
sured response factors for the main reaction products are: fcEtOH = 1,
fcAcH = 1.5, fcAcOEt = 2.1, fcOthers = 2.
represent respectively the numbers of car-
i
EtOH
i
2
2
−
1
of catalyst, has been taken in the range 1–20 g h mol , where W
and F are respectively the catalyst weight and the ethanol molar
flow rate. After the reachment of steady state conditions (about
3. Results and discussion
2
–3 h were necessary), the unreacted ethanol and the condens-
3.1. Catalytic results
able obtained products were collected, for about 30 min, in a trap
cooled with liquid nitrogen. The obtained reaction products were
injected and analyzed by a gas-chromatograph HP 6890, equipped
with a Restek Rt-Q Plot 30 m × 0.32 mm column, pure hydrogen
has been used as carrier gas. The conditions used for the analy-
Many different kinetic runs have been performed by chang-
ing the temperature, the pressure and the space time in the range
previously mentioned. In correspondence with each experimental
run different samples of the gaseous outlet mixture were with-
drawn, condensed, and analyzed by the GC method previously
described, to evaluate, when the reactor reached steady state
conditions, the ethanol conversion and products yields. For each
experiment, performed in steady-state conditions, three samples
have been analyzed and the results averaged. All the collected data
◦
ses were as follows: the initial temperature of 80 C was increased
◦
◦
at a rate of 10 C/min up to 220 C and then maintained at this
value for 10 min. The flame ionization detector (FID) was kept at
2
◦
◦
80 C. The split–splitless injector was kept at 250 C. Three differ-
ent sets of runs were performed in order to evaluate the effect of the

G. Carotenuto et al. / Catalysis Today 203 (2013) 202–210
205
Table 2
Operative conditions and experimental results for runs performed by chang-
ing the pressure and the temperature of reaction but keeping constant the
−
1
−2
ethanol space time W/FEtOH = 4 g h mol , nitrogen (5.7 × 10 mol/h) and hydrogen
−
3
(
4 × 10 mol/h) flow rates. 2 g of catalyst have been used in all the performed runs.
◦
T ( C)
P (atm)
XEtOH (%)
SACOEt (%)
SAcH (%)
SOthers (%)
200
200
200
220
220
220
240
240
240
260
260
260
200
200
200
220
220
220
240
240
240
260
260
260
200
200
200
220
220
220
240
240
240
260
260
260
10
10
10
10
10
10
10
10
10
10
10
10
20
20
20
20
20
20
20
20
20
20
20
20
30
30
30
30
30
30
30
30
30
30
30
30
16.16
15.47
16.79
25.94
24.55
23.07
39.80
38.59
40.62
46.46
42.18
47.26
10.60
10.89
10.24
25.33
25.27
24.29
32.08
31.99
31.74
42.79
39.76
42.18
17.30
16.94
17.03
20.89
20.30
22.64
25.47
23.60
23.21
35.35
28.76
29.40
62.79
59.38
63.16
67.60
66.99
66.94
70.64
68.21
71.78
70.21
70.84
70.33
81.07
85.73
82.96
79.18
80.19
77.59
78.37
78.70
77.00
80.81
80.29
80.54
76.02
77.68
79.85
80.23
80.08
84.41
70.36
70.39
70.58
73.74
72.14
70.55
37.21
40.62
36.84
26.35
27.02
27.09
21.10
24.55
22.22
21.84
21.80
21.61
18.93
14.27
17.04
16.85
16.46
18.93
17.27
16.94
18.25
12.14
13.59
13.32
20.9
18.58
16.85
15.63
15.20
12.12
25.02
23.15
24.89
19.45
22.17
23.49
0.01
0.01
0.01
6.05
5.99
5.97
8.26
7.24
6.00
7.95
7.36
8.06
0.01
0.01
0.01
3.97
3.35
3.48
4.36
4.36
4.75
7.05
6.12
6.14
3.10
3.70
3.30
4.12
4.72
3.46
1.38
6.46
4.53
6.81
5.69
5.96
were submitted to mathematical regression analysis. In Table 2,
the results obtained in the runs performed at different tem-
◦
peratures (200–260 C) and pressures (10–30 bar) but keeping
−
1
constant both the ethanol space time W/F = 4 g h mol
(ethanol
−
2
flow rate = 0.5 mol/h), the nitrogen (5.7 × 10 mol/h) and hydro-
−
3
gen (4 × 10 mol/h) flow rates, are reported.
In Table 3, the results obtained, by changing the ethanol
space time W/FEtOH from 1 to 20 g h mol , keeping constant the
temperature at 220 C, the pressure at 20 bar and the nitrogen
−1
◦
−
2
−3
(
5.7 × 10 mol/h) and hydrogen (4 × 10 mol/h) flow rates, are
summarized. At last, in Table 4, the results of the kinetic runs
◦
performed, at a constant temperature of 220 C and pressure of
2
0 bar, by changing the ethanol space time from 1 to 20 W/FEtOH
Table 3
Operative conditions and experimental results for runs performed by changing the
−
1
ethanol space time from 1 to 20 W/FEtOH (g h mol ), keeping constant the tempera-
ture at 220 ◦C, the pressure at 20 bar and nitrogen (5.7 × 10−2 mol/h) and hydrogen
−
3
(
4 × 10 mol/h) flow rates.
W/FEtOH
(
FEtOH
(mol/h)
XEtOH (%)
SACOEt (%)
SAcH (%)
SOthers (%)
g h mol−1)
1
1
4
4
4
1.5
1.5
0.5
0.5
0.5
0.1
0.1
13.74
13.63
25.33
25.27
24.29
46.02
47.35
53.61
53.13
79.18
80.19
77.59
90.62
90.25
40.37
40
16.85
16.46
18.93
5.21
6.02
6.87
3.97
3.35
3.48
4.17
4.16
2
2
0
0
5.59

2
06
G. Carotenuto et al. / Catalysis Today 203 (2013) 202–210
Table 4
Operative conditions and experimental results for runs performed by changing the ethanol space time from 1 to 20 W/FEtOH (g h mol−1), and the inlet hydrogen flow in the
−4
−3
◦
range 7.31 × 10 –3.66 × 10 mol/h, keeping constant the temperature at 220 C and the pressure at 20 bar.
W/FEtOH (g h mol 1
−
)
FN2 (mol/h)
FH2 (mol/h)
X (%)
SACOEt (%)
SAcH (%)
SOthers (%)
FEtOH (mol/h)
0.1
0.1
0.1
0.1
0.1
0.1
0.5
0.5
0.5
0.5
0.5
0.5
0.5
1.5
1.5
1.5
1.5
1.5
1.5
20
20
20
20
20
20
4
4
4
4
4
4
4
1
1
1
1
1
1
0.057
0.057
0.034
0.034
0.011
0.011
0.057
0.057
0.057
0.034
0.034
0.011
0.011
0.057
0.057
0.034
0.034
0.011
0.011
3.66E−03
3.66E−03
2.19E−03
2.19E−03
7.31E−04
7.31E−04
3.66E−03
3.66E−03
3.66E−03
2.19E−03
2.19E−03
7.31E−04
7.31E−04
3.66E−03
3.66E−03
2.19E−03
2.19E−03
7.31E−04
7.31E−04
46.02
47.35
46.77
46.77
33.26
34.33
25.33
25.27
24.29
19.09
19.27
16.90
15.79
13.74
13.63
8.70
90.62
90.25
84.17
83.07
75.90
72.50
79.18
80.19
77.59
63.96
63.99
61.97
63.58
53.61
53.13
46.66
44.41
37.81
36.46
5.21
5.59
4.17
4.16
3.96
3.58
6.86
5.81
3.97
3.35
3.48
6.66
5.72
5.42
5.36
6.02
6.87
–
11.87
13.35
17.24
21.69
16.85
16.46
18.93
29.38
30.29
32.60
31.06
40.37
40.00
53.34
55.59
62.19
63.54
8.85
8.41
8.74
–
–
–
−
1
−4
(
3
g h mol ) and the inlet hydrogen flow rate (from 7.31 × 10 to
3.3. Evaluation of the eventual external and internal diffusion
limitation
.66 × 10− mol/h) have been reported.
3
Intra-particles transport effect has been analyzed by applying
the Weisz and Prater criterium [15]. To ensure Á > 0.95 in an isother-
mal spherical particle for a first-order reaction, the criteria requires:
3.2. Reactor model
The collected kinetic data, at different pressures, tempera-
2
rR
p
tures and space times, were interpreted by a mono-dimensional
plug flow reactor model, for which an isothermal condition was
assumed. This assumption is justified by the relatively low ethanol
conversion obtained with 2 g of catalyst and by the high thermal
capacity of the reactor body. Considering a mass balance for each
component, this can be advantageously written in dimensionless
form by introducing the dimensionless length of the bed, z, defined
as z = L/LBED, as it follows:
< 1
(10)
C0D
eff
For power-law kinetics of different orders the criteria respectively
of Weisz [16] (11) and Hudgins [17] (12) the following relation can
be applied.
ꢋ
2
< 6 zero order
< 0.6 first order
rR
p
(11)
(12)
C0D
eff
<
0.3 second order
Nr
ꢋ
ꢅ
2
dFi
dz
rR
1
p
=
wcat
˛ (r )
(7)
<
ij
j
C0Deff
ꢁnꢂ
J=1
For a second order kinetics a limit of 0.3 can be adopted according to
Weisz criterium (11) and 0.5 according to Hudgins (12). Therefore,
In our experiments we can exclude the internal diffusive limita-
tions, because, we have found a maximum value of 0.15 for both
the mentioned criteria (11) and (12).
The system of ordinary differential equations (7) has been numer-
ically integrated using ODE45 function of MATLAB and a nonlinear
least squares fitting has been simultaneously performed for the
determination of the adjustable models parameters by using the
following objective function that was minimized by mathematical
regression analysis:
The average reaction rate r has been evaluated for all the per-
formed runs with the following relationship:
0
F XE
E
n
ꢅ
r =
(13)
RMS = 1
(ai − a ci alc)2
exp
(8)
(1 − ε )(w
B
cat
/ꢀB)
N
i=1
While the calculation of the effective diffusivity from molecular
diffusion coefficient, porosity, tortuosity and constriction factor has
been reported in Eq. (14).
In this relation aexp and acalc are respectively related to both exper-
imental and calculated conversions and selectivities. The overall
average error has been calculated by using the following expres-
sion:
ꢃ
ꢄ
ε × ꢂ
Deff = Di
(14)
ꢁ
By considering the properties reported in Table 5 and calculating
the ratio of the Weisz and Prater criterium for the highest observed
reaction rate we obtained a value of 0.15, that is, largely below the
limit of 1. In conclusion internal diffusion limitation is not oper-
ative. Then, eventual external mass transfer limitation has been
evaluated by adopting the Mears criterium [18].
ꢆ
ꢇ
ꢇ
ꢇ
ꢇ
ꢈ
ꢉꢇ
ꢇ
ꢊ
4
N
exp
ji
calc
ji
ꢅ
ꢅ
a
− a
1
ꢇ
err =
ꢇ100
(9)
exp
N ꢇ
a
ꢇ
ji
j=1
i=1
where j indicates an index for respectively ethanol conversion,
acetaldehyde selectivity, ethyl acetate selectivity and other by-
products selectivity.
rεbRpn
<
0.15
(15)
kcC0

G. Carotenuto et al. / Catalysis Today 203 (2013) 202–210
207
A 1.0
B 1.0
0.8
0.8
0
.6
0.6
0.4
0.2
0.0
0.4
Ethylacetate
Acetaldehyde
Others
0.2
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Experimental conversion
Experimental selectivities
Fig. 1. (A and B) Parity plots related to respectively the ethanol conversion and selectivities by considering a catalyst bed of 2 g.
The mass transfer coefficient kc appearing in Eq. (15) has been eval-
uated by using the Thoenes–Kramers correlation [19]:
for determining the equilibrium compositions and then the cor-
responding gas phase equilibrium constants have been estimated.
Both the effects of pressure and of the non-ideality of the gaseous
mixtures on equilibrium have been evaluated by adopting the PSRK
equation of state. Our kinetic runs have been performed in a nar-
row range of temperature, therefore, we have applied the Van’t
Hoff equation for evaluating the dependence of the equilibrium
constants by the temperature (20).
ꢀ
ꢂ
1
− εB
Di
dp
ꢀ
kc =
Sh
(16)
εB
As in the previous case, in correspondence of the highest average
reaction rates, we calculated a value for the expression (16) of less
than 0.1, that is, also in this case, we can conclude that the external
mass transfer has not influence. By concluding, our runs have been
performed in chemical regime.
ꢌ
ꢀ
ꢂꢍ
ꢃ
ꢄ
ꢅHi
R
1
Tref
1
Ke = K
× exp
− T
(20)
i
eref
The equilibrium constants have been estimated determining the
3.4. Equilibrium constants estimation
outlet composition of an equilibrium reactor in ASPEN by consid-
ering the two main reactions, one related to acetaldehyde (Ke ) and
The main occurring equilibrium reactions and the correspon-
dent enthalpies are:
1
the other one to ethylacetate (Ke ) formation. The dependence of
2
the equilibrium constants on the temperature has been expressed
by the following Eq. (21):
CH CH OH ↔ CH CHO + H
16.45 kcal/mol
(17)
3
2
3
2
B
ln Ke = A +
(21)
i
T
CH CH OH + CH CHO ↔ CH COOCH CH + H
3
2
3
3
2
3
2
The values of the constants A and B are the following: reaction
(17) A = 16.5, B = −9136.4, reaction (18) A = −4.79, B = 4386.0. These
expressions have directly been used in the developed kinetic model.
−10.37 kcal/mol
(18)
(19)
By adding the two reactions we obtain the overall reaction:
2
CH CH OH ↔ CH COOCH CH + 2H 5.98 kcal/mol
3.5. Kinetic approach and reaction mechanism
3
2
3
2
3
2
As it can be seen, the overall reaction (19) is moderately endother-
mic and this justifies the assumption of isothermicity for a short
catalytic bed. To evaluate the equilibrium constants of reaction (17)
and (18) the method of Gibbs energy minimization has been used
The kinetic approach has been developed on the following reac-
tion scheme, considering three different main reactions:
CH CH OH → CH CHO + H
(22)
(23)
(24)
3
2
3
2
CH CH OH + CH CHO → CH COOCH CH + H
3
2
3
3
2
3
2
Table 5
Properties used for evaluating internal and external diffusion effects.
2CH3CHO → otherproducts
Properties
Symbol
Value
A first kinetic evaluation was done by using the following empirical
power laws:
Bed porosity
Bulk density
Tortuosity
Constriction factor
Molecular diffusivity EtOH in N2
Pellets diameter
Cross sectional area of the tube
Kinematic viscosity
Assumed reaction order
εB
0.4
0.95
0.4
3
ꢈ
ꢉ
ꢀB (g/cm )
ꢁ
ˇ
ꢆ
P
P
1
ke1
AcH H2
˛
ꢂ
0.8
r1 = k1PEtOH 1 −
˛
(25)
(26)
P
2
EtOH
Di (cm /min)
6.12
0.26
2.54
0.0092
2
dp (cm)
ꢈ
ꢉ
2
ꢇ
ω
Ac (cm )
P
P
AcH H2
1
2
ı
ε
ꢄ (cm /min)
r2 = k2P
P
1 −
ke2
EtOH AcH
ı
ε
n
P
P
EtOH
A

2
08
G. Carotenuto et al. / Catalysis Today 203 (2013) 202–210
A 1.0
0.0
0.2
0.4
0.6
0.8
1.0
1.0
B 1.0
0
.8
0
.8
0.8
0.6
0.4
0.2
0.0
0.6
0.6
0.4
0.4
Ethylacetate
Acetaldehyde
Others
0.2
0.2
0.0
0.0
0
.0
0.2
0.4
0.6
0.8
1.0
0.0
0.2
0.4
0.6
0.8
1.0
Experimental conversion
Experimental selectivities
Fig. 2. (A and B) Parity plots related respectively to the ethanol conversion and selectivities by considering a catalyst bed of 50 g.
2
r3 = k3P
(27)
rate laws. The best performances have been obtained by consid-
ering as RDS for respectively the formation of acetaldehyde, ethyl
acetate and others by-products, the three reactions (28.c), (29.c)
and (30.b). This dual-site model has shown satisfactory perfor-
mances in terms of average error (17%) but some of the parameters
found were not meaningful. In particular, some activation ener-
gies resulted close to zero. Moreover, this model and all the others,
based on the previously reported mechanisms, were not able to
describe with a good accuracy the experimental data collected by
using 50 g of catalyst (close to equilibrium). Then, we tested the
performances of a LHHW model, based on the molecular associative
hydrogen adsorption mechanism.
AcH
The best values obtained for the reaction orders ˛, ˇ, ε, ꢈ, ꢇ, ı, ꢆ
suggested us a reaction mechanism in which the adsorption of both
reactants and products plays an important role, that is, the power-
law model suggested kinetic expressions with the partial pressures
of different species at the denominator.
According to the literature the adsorption of hydrogen on
copper surface can occur by both an associative and/or a dis-
sociative mechanism [20,21]. In the case associative adsorption,
molecular hydrogen adsorbed on a single atom gives place
to a weakened hydrogen–hydrogen bond and adsorbed hydro-
gen is more reactive. According to some authors [22] also the
dissociative mechanism on two adjacent sites is possible, as nor-
mally occurs for other metals (palladium, platinum etc.). On the
basis of these considerations, two models have been developed,
one based on the dissociative adsorption (dual site mechanism
DSM) and the other based on the molecular associative one
The acetaldehyde formation is described through the following
elementary steps:
C2H5OH + ꢂ0 ⇔ ꢂEtOH
(31.a)
(31.b)
(31.c)
(31.d)
ꢂEtOH + ꢂ0 → ꢂAcH + ꢂH2 rate determining step
(
Langmuir–Hinshelwood–Hougen–Watson; LHHW). For example,
the dual sites model, involving two catalytic sites, is based on a
mechanism for the acetaldehyde formation, as it follows:
ꢂAcH ⇔ ꢂ0 + CH CHO
3
ꢂH2 → ꢂ + H2
0
Then, for the reaction to ethyl acetate we suggest the following
mechanism
CH CH OH(g) + 2ꢂ ⇔ CH CH O(a) + H(a)
(28.a)
(28.b)
(28.c)
(28.d)
3
2
0
3
2
CH CH O(a) + ꢂ ⇔ CH CHO(a) + H(a)
3
2
0
3
C2H5OH + ꢂ0 ⇔ ꢂEtOH
(32.a)
(32.b)
(32.c)
(32.d)
(32.e)
CH CHO(a) ⇔ CH CHO(g) + ꢂ
3
3
0
CH CHO + ꢂ0 ⇔ ꢂAcH
3
2
H(a) ⇔ H + 2ꢂ
2
0
ꢂEtOH + ꢂAcH ⇔ ꢂEA + ꢂH2 rate determining step
ꢂEA ⇔ꢂ0 + CH COOC H5
A similar approach can be developed also for the mechanism of
ethyl acetate formation.
3
2
ꢂH2 ⇔ ꢂ + H2
0
CH CHO(a) + ꢂ ⇔ CH CO(a) + H(a)
(29.a)
(29.b)
(29.c)
3
0
3
In this case, the rate determining step should be the dual sites reac-
tion between two adsorbed molecules of respectively ethanol and
acetaldehyde to give adsorbed ethyl acetate and hydrogen.
The third reaction to other by-products (24), occurring in a very
small amount, between two molecules of adsorbed acetaldehyde
to give other by-products, has been described in an approximated
way by a second order not reversible reaction. Therefore, on the
basis of the described mechanisms, the following kinetic rate laws
can be derived (33)–(35):
CH CO(a) + CH CH O(a) ⇔ CH COOC H5(a)
3
3
2
3
2
CH COOC H5(a) ⇔ CH COOC H5(g) + 2ꢂ
3
2
3
2
0
At last, the by-products formation can be described through the
following elementary reactions:
CH CHO(a) + CH CHO(a) ⇔ CH CH(OH)CH CHO(a)
(30.a)
(30.b)
3
3
3
2
CH CH(OH)CH CHO(a) ⇔ CH CH(OH)CH CHO(g) + 2ꢂ
3
2
3
2
0
By considering the two first mechanisms, i.e., the formation of
respectively acetaldehyde and ethyl acetate, different rate deter-
mining steps (RDS) have been considered giving place to different
k bEtOHPEtOH(1 − (1/Ke )(P
PH /PEtOH))
2
1 + bEtOHPEtOH + bAcHPAcH + bHPH + bEAPEA)2
1
1
AcH
r1
=
(33)
(

G. Carotenuto et al. / Catalysis Today 203 (2013) 202–210
209
Table 6
all the catalytic sites are initially occupied by adsorbed hydrogen.
Moreover, hydrogen is fed with the reagents. For these reasons a
correct evaluation of this parameter is difficult in the described con-
ditions. On the other hand, the co-feeding of hydrogen during the
reaction is fundamental for limiting the acetaldehyde formation
that is a prerequisite to increase the selectivity to ethyl acetate and
preventing the acetaldehyde condensation to by products. More-
over, the presence of hydrogen prevents the catalyst fouling and
consequent deactivation during the reaction [11]. Then, it is known,
from the literature, that the presence of hydrogen is beneficial pre-
venting the copper sintering [23]. Another unsuitable parameter is
the activation energy of the reaction giving by-products. It is clear
that in this case the assumption of second order kinetics and the
difficulty of analytical response for very low concentrations of these
substances are responsible of the poor result.
Kinetic parameters of the LHHW dual site model determined by regression analysis
on all the experimental runs performed by using the reactor loaded with 2 g of
catalyst.
Kinetic constants
Activation energy
(
kcal/mol)
k1-220 ◦ C
k2-220 ◦ C
k3-220 ◦ C
97.1 ± 6.8 (mol/(gcat h atm))
36.25 ± 4.35
12.95 ± 0.65
2
0.089 ± 9.8E−3 (mol/(gcat h atm ))
2
0.0011 ± 7.8E−4 (mol/(gcat h atm ))
1.6E−4 ± 1.8E−05
Adsorption parameters
Adsorption enthalpy
kcal/mol)
(
10.4 ± 0.83 (atm−
1
)
−25.53 ± 2.55
−7.02 ± 0.35
−13.91 ± 0.14
−13.34 ± 1.47
bEtOH-220 ◦ C
bAcH-220 ◦ C
bEA-220 ◦ C
bH-220 ◦ C
98.4 ± 12.79 (atm 1)
−
41.2 ± 4.94 (atm−1)
2.5 × 10−4 ± 3.50E−5 (atm
−1
)
However, the average standard error resulted less than 15%
and the validity of the developed model can be appreciated in
Fig. 1A and B in which are reported the parity plots for respectively
the ethanol conversion and the selectivity to acetaldehyde, ethyl
acetate and others.
k bEtOHbAcHPEtOHPAcH(1 − (1/Ke )(P PH /PEtOHPAcH))
2
2
EA
2
r2
=
(34)
(35)
1 + bEtOHPEtOH + bAcHPAcH + bHPH + bEAPEA)2
(
2
AcH
r3 = k3P
In Table 6, all the kinetic parameters determined by regression anal-
ysis are reported. The equilibrium constants Ke1 and Ke2 have been
calculated, as described in Section 3.4. From the kinetic parame-
ters reported in Table 6 it is possible to observe that k3 has a very
low value and a negligible value of the activation energy. This is
the consequence of: (i) the approximation introduced by consider-
ing a pseudo-second order rate law; (ii) the fact that reaction (3)
corresponds to an ensemble of different reactions considered as
acetaldehyde that gives “others”; (iii) the very low amount of by-
products found corresponding to a low precision in the analytical
determination.
3.6. Use of the kinetic model for the simulation of other kinetic
runs
As mentioned in a previous section, some kinetic runs have been
performed by charging the same tubular reactor with 50 g of cat-
alyst instead of 2 g. In this case, two different consecutive ovens
have been used for heating the reactor with the aim to obtain as
much as possible an isothermal profile along the catalytic bed. This
has been experimentally verified and a variation of not more than
◦
The parameters values are acceptable with the exclusion of the
parameter related to the adsorption of hydrogen. However, the
obtained value can be justified taking into account the fact that the
catalyst is pre-reduced with hydrogen and very probably almost
± 5 C has been observed along the reactor. In Table 7 some of the
performed runs are reported together with the adopted experimen-
tal conditions. In the same table are also reported for comparison
experimental and calculated values of ethanol conversions and
Table 7
◦
Comparison of experimental and calculated data by using LHHW model. The runs have been performed by using 50 g of catalyst, temperature range: 200–260 C, pressure
range: 10–30 bar, at two different space time of 34.4–97.32 g h mol , keeping constant the nitrogen (5.7 × 10 mol/h) and hydrogen (4 × 10 mol/h) flow rates.
−
1
−2
−3
Operative conditions
Experimental data
Calculated data with the LHHW model
◦
3
−1
W/F (g h mol )
T ( C)
P (bar)
FEtOH (cm /min)
XEtOH
SAcOEt
SAcH
SOthers
XEtOH
SAcOEt
SAcH
SOthers
200
220
240
260
200
200
220
220
240
240
260
200
220
220
240
260
200
220
240
260
200
220
240
260
200
220
240
240
10
10
10
10
20
20
20
20
20
20
20
20
30
30
30
30
10
10
10
10
20
20
20
20
30
30
30
30
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
97.32
97.32
97.32
97.32
97.32
97.32
97.32
97.32
97.32
97.32
97.32
97.32
97.32
97.32
97.32
97.32
34.40
34.40
34.40
34.40
34.40
34.40
34.40
34.40
34.40
34.40
34.40
34.40
45.26
51.14
61.8
96.19
96.1
96.53
93.41
98.79
98.3
3.18
2.9
0.63
1
2.36
5.07
0
0.42
0.01
0.01
0
50.60
63.28
70.13
74.00
43.73
43.73
55.68
55.68
63.29
63.29
67.93
43.73
51.19
51.19
59.16
64.34
35.90
50.43
63.84
72.42
30.86
43.26
55.78
65.33
27.85
39.10
51.01
51.01
97.27
96.74
95.33
92.78
96.97
96.97
95.99
95.99
94.03
94.03
90.58
96.97
94.99
94.99
92.36
87.80
94.78
94.81
94.44
93.13
95.01
94.31
93.22
91.28
94.55
93.29
91.51
91.51
1.74
1.91
2.68
4.16
1.29
1.29
1.44
1.44
1.97
1.97
2.97
1.29
1.23
1.23
1.64
2.42
3.98
3.66
3.64
4.36
2.82
2.75
2.81
3.27
2.36
2.36
2.43
2.43
0.99
1.35
1.99
3.06
1.74
1.74
2.57
2.57
4.00
4.00
6.45
1.74
3.78
3.78
6.00
9.79
1.24
1.53
1.92
2.51
2.17
2.94
3.97
5.45
3.08
4.36
6.06
6.06
1.11
1.52
1.21
1.28
0.66
1.23
0.42
0.47
0.52
69.7
49.58
48.63
57.33
60.54
64.83
64.33
70.63
54.29
61.3
63.21
63.21
67.17
34.81
40.5
59.94
60.38
35.32
47.21
61.17
67.54
35.93
42.37
60.87
59.19
99.33
98.76
99.58
93.43
94.25
98.95
98.94
96.84
96.84
96.04
78.18
86.91
84.05
84.18
93.53
96.91
96.1
6.1
5.23
1.05
0.67
2.57
2.6
3.16
11.24
6.78
9.65
10.94
1.26
1.17
2.1
11
0.39
0.59
0.56
0.8
10.58
6.31
6.3
4.88
5.21
1.92
1.8
1.75
3.68
1.99
2.11
1.47
95.23
94.79
96.08
95.8
3.02
1.53
1.93
2.09
1.87
96.66

2
10
G. Carotenuto et al. / Catalysis Today 203 (2013) 202–210
1
,0
,9
,8
,7
,6
,5
,4
,3
,2
,1
,0
ethanol to be added to the ethanol recycle stream. A purge would
be sufficient to maintain steady state conditions in the reactor.
In this paper, the kinetics of the occurring reactions have been
studied for the scope of reactor modeling and optimization. A
Langmuir–Hinshelwood–Hougen–Watson kinetic model has been
used for interpreting all the kinetic runs performed, that is, 62 runs
performed in different operative conditions by using a tubular reac-
tor filled with 2 g of catalyst and 28 runs made by using 50 g of
catalyst. It has been shown that the runs with the lowest amount
of catalyst have been performed in chemical regime and have been
used to identify the best kinetic model, while, the runs performed
with 50 g of catalyst give data that are near the equilibrium condi-
tions and allow to verify both the model goodness and the validity
of the equilibrium constants. The obtained agreements are satisfac-
tory, considering the approximations introduced as the assumption
of isothermal condition and the use of the equilibrium constants
derived from theoretical calculations. At last, the model is based
on a reliable mechanism and the kinetic parameters show physical
mean.
0
0
2g
50g
0
0
0
0
LHHW Model
SAcH
0
SAE
0
XEtOH
0
0
0
20
40
60
-1
80
100
W/F (ghmol )
Fig. 3. Conversion and selectivities obtained for different space times. This plot
has been obtained by considering all data collected in both the reactors contain-
References
◦
ing respectively 2 and 50 g of catalyst working at 220 C, 20 bar with a constant flow
3
of a mixture of 6% H2 in N2 of 25 cm /min that correspond to an hydrogen flow of
[
1] B.N. Dolgov, M.M. Koton, N.V. Siderov, Journal of General Chemistry of the USSR
6 (1936) 1456.
3
3
.77 × 10− mol/h and nitrogen flow of 0.057 mol/h.
[
[
2] Y.J. Tu, Y.W. Chen, C. Li, Journal of Molecular Catalysis 89 (1994) 179–189.
3] Y. Tu, J. Li, C.Y.W. Chen, Journal of Chemical Technology and Biotechnology 59
selectivities of respectively ethyl acetate, acetaldehyde and other
by-products.
(
1994) 141–147.
[4] S.W. Colley, J. Tabatabaei, K.C. Waugh, M.A. Wood, Journal of Catalysis 236
2005) 21–33.
5] N. Iwasa, N. Takezawa, Bulletin of the Chemical Society of Japan 64 (1991)
619–2623.
[6] D.J. Elliot, E. Pennella, Journal of Catalysis 119 (2) (1989) 359–376.
(
Calculations have been made with the LHHW described model
using the kinetic parameters reported in Table 6 and the agree-
ment obtained is very satisfactory as it can be appreciated in Table 7
where experimental and calculated data are reported for compar-
ison. In Fig. 2A and B are reported the parity plots of respectively
conversions and selectivities for the runs performed with 50 g of
catalyst.
However, the results obtained in these runs correspond, very
probably, to equilibrium conditions. This can be appreciated in
Fig. 3, in which the profiles of respectively conversions and selec-
tivities are reported as a function of the space time. As it can be
seen, the runs made with 50 g of catalyst show the approaching to
a plateau for both conversion and selectivities. For this reason these
runs have not been considered together with the ones performed
with 2 g of catalyst in the regression analysis but are used here to
verify the model. The average standard error in simulating these
runs is about 12%.
[
2
[
7] K. Inui, T. Kurabayashi, S. Sato, Applied Catalysis A: General 237 (2002)
3–61.
5
[
8] K. Inui, T. Kurabayashi, S. Sato, N. Ichikawa, Journal of Molecular Catalysis A:
General 216 (2004) 147–156.
[9] K. Inui, T. Kurabayashi, S. Sato, Journal of Catalysis 212 (2002) 207–215.
10] A.B. Gaspar, F.G. Barbosa, S. Letichevsky, L.G. Appel, Applied Catalysis A: General
[
[
[
[
380 (2010) 113–117.
11] E. Santacesaria, G. Carotenuto, R. Tesser, M. Di Serio, Chemical Engineering
Journal 179 (2012) 209–220.
12] S.W. Colley, C.R. Fawcett, M. Sharich, M. Tuck, D.J. Watson, M.A. Wood, US Patent
7,553,397, B1 June 30, 2009.
13] G. Carotenuto, M. Di Serio, E. Santacesaria, R. Tesser, New process for the
production of ethylacetate and pure hydrogen from ethanol, Italian Patent
NA2010A000009, 2010; WO2011/104738A2 assigned to EUROCHEM Engineer-
ing.
[
14] A.A. Khasin, T.M. Yur’eva, L.M. Plyasova, G.N. Kustova, H. Jobic, A. Ivanov, Yu.A.
Chesalov, V.I. Zaikovskii, A.V. Khasin, L.P. Davydova, V.N. Parmon, Russian Jour-
nal of General Chemistry 78 (11) (2008) 2203–2213.
[
[
15] P.B. Weisz, C.B. Prater, Advances in Catalysis 6 (1954) 143.
16] P.H. Weisz, Zeitschrift fur Physikalische Chemie Neue Folge 11 (1957) 1.
4
. Conclusions
[17] R.R. Hudgins, Chemical Engineering Science 23 (1968) 93.
[
18] D.E. Mears, Industrial & Engineering Chemistry Process Design and Develop-
ment 10 (4) (1971).
In a previous work [11] we have shown that the ethanol dehy-
[
19] J.B. Butt, Reaction Kinetics and Reactors Design, 2001.
drogenation to ethyl acetate occurs on a copper/copper chromite
catalyst, supported on alumina and promoted with barium chro-
mate, with a conversion of 65–70% and a selectivity of 98–99% in the
[20] A.I. Seryhk, V.B. Kazansky, Physical Chemistry Chemical Physics 6 (2004)
250–5255.
21] X.-J. Kuang, X.-Q. Wang, G.B. Liu, Journal of Chemical Sciences 123 (September
5)) (2011) 743–754.
[22] L. Triguero, U. Wahlgren, P. Boussard, P. Siegbahn, Chemical Physics Letters 237
(1995) 550.
5
[
(
◦
−1
optimal conditions that are: 220 C, 20 bar and W/F = 100 g h mol
.
A so high selectivity suggests the use of the mentioned catalyst
in a new competitive process not requiring any post-treatment
as for example the hydrogenation of co-produced acetaldehyde to
[
23] J.R. Lewis, H.S. Taylor, Journal of the American Chemical Society 60 (4) (1938)
77–879.
8